lot1-kickoff/airflow/dags/import_EOSC_graph.py

from __future__ import annotations

import gzip
import io
import json
import logging
import os
from datetime import timedelta

from kubernetes.client import models as k8s
import pendulum
from airflow.decorators import dag
from airflow.decorators import task
from airflow.operators.python import PythonOperator
from airflow.providers.amazon.aws.hooks.s3 import S3Hook
from airflow.utils.helpers import chain
from airflow.hooks.base import BaseHook

from opensearchpy import OpenSearch, helpers
from EOSC_indexes import mappings
from EOSC_entity_trasform import transform_entities


S3_CONN_ID = os.getenv("S3_CONN_ID", "s3_conn")
EXECUTION_TIMEOUT = int(os.getenv("EXECUTION_TIMEOUT", 6))

ENTITIES = ["datasource", "grants", "organizations", "persons", "products", "topics", "venues"]

BULK_PARALLELISM = 10

default_args = {
    "execution_timeout": timedelta(days=EXECUTION_TIMEOUT),
    "retries": int(os.getenv("DEFAULT_TASK_RETRIES", 1)),
    "retry_delay": timedelta(seconds=int(os.getenv("DEFAULT_RETRY_DELAY_SECONDS", 60))),
}

@dag(
    schedule=None,
    dagrun_timeout=None,
    start_date=pendulum.datetime(2021, 1, 1, tz="UTC"),
    catchup=False,
    default_args=default_args,
    params={
        "S3_CONN_ID": "s3_conn",
        "OPENSEARCH_CONN_ID": "opensearch_default",
        "EOSC_CATALOG_BUCKET": "eosc-portal-import"
    },
    tags=["lot1"]
)
def import_EOSC_graph():
    @task
    def create_indexes(**kwargs):
        conn = BaseHook.get_connection(kwargs["params"]["OPENSEARCH_CONN_ID"])
        client = OpenSearch(
            hosts=[{'host': conn.host, 'port': conn.port}],
            http_auth=(conn.login, conn.password),
            use_ssl=True,
            verify_certs=False,
            ssl_show_warn=False,
            pool_maxsize=20
        )

        client.cluster.put_settings(body={
            "persistent": {
                "cluster.routing.allocation.balance.prefer_primary": True,
                "segrep.pressure.enabled": True
            }
        })

        for entity in ENTITIES:
            if client.indices.exists(entity):
                client.indices.delete(entity)

            client.indices.create(entity, {
                "settings": {
                    "index": {
                        "number_of_shards": 40,
                        "number_of_replicas": 0,
                        "refresh_interval": -1,

                        "translog.flush_threshold_size": "2048MB",

                        "codec": "zstd_no_dict",
                        "replication.type": "SEGMENT"
                    }

                },
                "mappings": mappings[entity]
            })

    def compute_batches(ds=None, **kwargs):
        hook = S3Hook(S3_CONN_ID, transfer_config_args={'use_threads': False})
        pieces = []
        for entity in ENTITIES:
            keys = hook.list_keys(bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"], prefix=f'{entity}/')
            to_delete = list(filter(lambda key: key.endswith('.PROCESSED'), keys))
            for obj in to_delete:
                hook.get_conn().delete_object(Bucket=kwargs["params"]["EOSC_CATALOG_BUCKET"], Key=obj)
            for key in keys:
                if key.endswith('.gz'):
                    pieces.append((entity, key))

        def split_list(list_a, chunk_size):
            for i in range(0, len(list_a), chunk_size):
                yield {"files": list_a[i:i + chunk_size]}

        return list(split_list(pieces, len(pieces) // BULK_PARALLELISM))

    @task(executor_config={
        "pod_override": k8s.V1Pod(
            spec=k8s.V1PodSpec(
                containers=[
                    k8s.V1Container(
                        name="base",
                        resources=k8s.V1ResourceRequirements(
                            requests={
                                "cpu": "550m",
                                "memory": "256Mi"
                            }
                        )
                    )
                ]
            )
        )
    })
    def bulk_load(files: list[(str, str)], **kwargs):
        conn = BaseHook.get_connection(kwargs["params"]["OPENSEARCH_CONN_ID"])
        client = OpenSearch(
            hosts=[{'host': conn.host, 'port': conn.port}],
            http_auth=(conn.login, conn.password),
            use_ssl=True,
            verify_certs=False,
            ssl_show_warn=False,
            pool_maxsize=20
        )
        hook = S3Hook(S3_CONN_ID, transfer_config_args={'use_threads': False})

        for (entity, key) in files:
            if hook.check_for_key(key=f"{key}.PROCESSED", bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"]):
                print(f'Skipping {entity}: {key}')
                continue
            print(f'Processing {entity}: {key}')
            s3_obj = hook.get_key(key, bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"])
            with s3_obj.get()["Body"] as body:
                with gzip.GzipFile(fileobj=body) as gzipfile:
                    def _generate_data():
                        buff = io.BufferedReader(gzipfile)
                        for line in buff:
                            data = json.loads(line)
                            data['_index'] = entity
                            data['_id'] = data['local_identifier']
                            if entity in transform_entities:
                                data = transform_entities[entity](data)
                            yield data

                    # disable success post logging
                    logging.getLogger("opensearch").setLevel(logging.WARN)
                    succeeded = 0
                    failed = 0
                    for success, item in helpers.parallel_bulk(client, actions=_generate_data(),
                                                               raise_on_exception=False,
                                                               raise_on_error=False,
                                                               chunk_size=5000,
                                                               max_chunk_bytes=50 * 1024 * 1024,
                                                               timeout=180):
                        if success:
                            succeeded = succeeded + 1
                        else:
                            print(item["index"]["error"])
                            failed = failed + 1

                    if failed > 0:
                        print(f"There were {failed} errors:")
                    else:
                        hook.load_string(
                            "",
                            f"{key}.PROCESSED",
                            bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"],
                            replace=False
                        )

                    if succeeded > 0:
                        print(f"Bulk-inserted {succeeded} items (streaming_bulk).")

    @task
    def close_indexes(**kwargs):
        conn = BaseHook.get_connection(kwargs["params"]["OPENSEARCH_CONN_ID"])
        client = OpenSearch(
            hosts=[{'host': conn.host, 'port': conn.port}],
            http_auth=(conn.login, conn.password),
            use_ssl=True,
            verify_certs=False,
            ssl_show_warn=False,
            pool_maxsize=20,
            timeout=180
        )
        for entity in ENTITIES:
            client.indices.refresh(entity)

    parallel_batches = PythonOperator(task_id="compute_parallel_batches", python_callable=compute_batches)

    chain(
        create_indexes.override(task_id="create_indexes")(),
        parallel_batches,
        bulk_load.expand_kwargs(parallel_batches.output),
        close_indexes.override(task_id="close_indexes")()
    )


import_EOSC_graph()